Project Summary: Cancer is a major cause of death worldwide with outstanding challenges for a cure. Such challenges are primarily due to the nature of tumor heterogeneity and evolvability. Thus, the ability to generate unbiased, quantitative and causal maps of functional drivers and their combinations in native tumor microenvironment is a key to accelerate therapeutic discovery. To date, little has been done to comprehensively and combinatorially test which of the mutations identified in human patients can indeed functionally drive tumorigenesis of normal cells in native organs. The major barriers include accurate delivery, precise genome manipulation, efficient massively parallel perturbation, and unbiased, high-sensitivity quantitative readout, all of which have to be achieved simultaneously in the native tissue microenvironment. We recently established a novel approach named Pooled AAV Screen with Targeted Amplicon Sequencing (PASTAS) for direct in vivo screening of causative cancer drivers and combinations. This method generates precision models of cancer that (1) spontaneously develop from tumor-originating cells in the native organ microenvironment, (2) develop in fully immunocompetent animals and preserve the immune microenvironment, (3) genetically mimic significant mutations found in patients, (4) closely mimic the histopathology of human disease and clinical features, (5) encompass high degree of genetic and cellular heterogeneity, (6) offer flexibility to target any choice of target genes and rapidly scalable as pooled mutant screens, and (7) is easy to use by the community. In this study, we will conduct advanced development, robust validation and full establishment of this screening system. We will first establish technical parameters for optimal performance of this technology by quantitative measurements using independent patient cohorts with two lethal cancer types: glioblastoma and liver hepatocellular carcinoma. Then, we will extend the utility for causative driver discovery in therapeutic settings. Finally, we will advance the development of a lentiviral vector-based orthogonal approach to open up larger screening capabilities. Such screening systems and models will enable rapid identification of causative factors that directly drive transformation of healthy cells, tumor initiation, progression and therapeutic responses to treatments. More importantly, compared to existing alternatives, the fully immunocompetent setting allows robust pre-clinical testing of immunotherapies, in genetically matched animal avatars, as well as screening for genes that modulate the response to these therapies. Outcome and impact: This R33 will deliver optimized and validated PASTAS / PLeSTASS systems to link causative genes to oncogenesis in native TME; to enable autochthonous immunotherapy screen in fully immunocompetent setting for identification of targets that modulate the response to these life-saving drugs; and to share resources and protocols for the community to collectively yield novel and far-reaching insights in oncology.